Dongjin Semichem (005290): A Korean Photoresist Maker Is Selling Its China Business – and Samsung Is Most of What’s Left

Almost every DRAM and NAND chip that leaves a Korean fab passes through a layer of light-sensitive polymer, and a large share of that polymer is mixed in Incheon by a company with no English Wikipedia page. Dongjin Semichem has supplied photoresist to Samsung Electronics since 1989. In its latest filing, Samsung is 62.1% of what it sells. A year earlier that figure looked completely different — not because Samsung bought more, but because Dongjin is selling the business that sold to everyone else.

🏢 What Is Dongjin Semichem, and What Does It Actually Sell?

Dongjin Semichem is a Korean specialty-chemicals company that makes photoresist — the light-sensitive polymer that lets a chipmaker print circuit patterns onto a wafer — plus the strippers, etchants and cleaning chemicals used in the same lithography step.

The company’s own description of itself is unusually plain. From the FY2025 annual report: “The company was established in 1967, developed and localised PVC and rubber foaming agents for the first time in Korea… entered the semiconductor and display materials industry in the early 1980s… and in 1989 developed photoresist for semiconductors in-house.” (Source: DART annual report FY2025, rcpNo 20260319000264, filed 19 Mar 2026.) Every quotation from a filing in this article is our own translation from the Korean original; none of these documents is published in English.

Two founding dates appear in the filings and they are not a contradiction: the business dates to 1967, the legal entity to July 1973. Headquarters are at 644 Baekbeom-ro, Seo-gu, Incheon, with plants in Hwaseong, Siheung and Eumseong and overseas entities in China, Taiwan, Sweden, the United States and — since May 2026 — Japan. The parent company employed 1,369 people at 30 June 2026, average tenure 8.8 years. Chairman and CEO Lee Jun-hyuk — Seoul National University chemical engineering, PhD from MIT — has been with the company 29 years.

The product line, in the filing’s own words: photoresist, BARC (bottom anti-reflective coating), spin-on-carbon, CMP slurry, wet chemicals, colored resist, organic insulating layers and column spacers. The brands are DPR and DNR in electronic materials and UNICELL in foaming agents. R&D ran ₩37,645m ($28.0M) in H1 2026, or 5.58% of revenue.

📊 Where Does Dongjin’s Revenue Actually Come From?

Two-thirds of it is semiconductor materials, a quarter is display materials, and the battery business everyone talks about is 1.5%.

There is a second cut of the same ₩674,960m that is worth reading slowly, because it inverts what the labels suggest. Dongjin books 78.17% of revenue as “export” and 21.83% as “domestic.” But the filing prints the sales channel behind each line, and the single biggest export line — ₩499,404m, 73.99% of revenue — is described as: “orders from Samsung Electronics, LG Display and others → Dongjin delivers and negotiates → Samsung Electronics, LG Display purchase-approval certificate → settlement.”

That is a local export: goods sold to a Korean company against a purchase-approval certificate, which Korean practice treats as an export for tax and customs purposes. Meanwhile the line that genuinely happens offshore — ₩82,217m (12.18%), described as “overseas affiliate produces and delivers → overseas customer settles” — is booked as domestic, because from the consolidated group’s viewpoint nothing crossed a Korean border. The headline export ratio and the geographic reality point in opposite directions.

🗓️ How Did Dongjin Semichem Get Here?

Nearly sixty years of moving up the value chain — from shoe-sole foaming agents to lithography chemicals — followed by an abrupt reversal in 2025, when the company decided to sell ten Chinese subsidiaries at once.

🌍 Is Dongjin a Global Leader in Photoresist?

We could not verify a market-share figure, because Dongjin does not publish one: the word “share” (점유) appears zero times in both its FY2025 annual report and its H1 2026 half-year report.

We checked two documents, not one screen. In the full text of both the FY2025 annual report and the H1 2026 half-year report, the Korean term for market share does not appear once. That is unusual — Korean issuers routinely disclose share estimates, and every other Korean company deep-dive we have written had at least one. So the honest statement is: we did not find one, not there isn’t one.

What the filings do assert, and what we can therefore only report as management’s characterisation:

• The company describes itself as maintaining “an industry-leading position” on the strength of its R&D — company-reported, unquantified.
• On foaming agents it claims its Microsphere product is “the world’s fourth commercialised, after Sweden and Japan” — company-reported, and note that this business was spun into Dongjin Innochem on 1 January 2026.
• On competitive structure the filing is candid rather than promotional: “the principal customers for semiconductor materials are two companies, Samsung Electronics and hynix… the domestic market’s main customers consist of three to five companies, and it is a fact that domestic demand alone cannot be expected to support revenue beyond a certain level.”

That last sentence is management describing its own ceiling. It is also, read alongside the customer table below, the most useful thing in the document.

🔍 Who Actually Buys Dongjin’s Photoresist?

Samsung Electronics alone was ₩419,197m — 62.11% of continuing-operations revenue — in the six months to June 2026, and Samsung plus SK hynix together account for the entire “customers above 10%” figure in the accounting notes.

This is worth walking through, because the disclosure comes in two places that do not obviously connect.

The IFRS revenue note (note 2.4, consolidated) is deliberately anonymous: “Of goods-sales revenue of ₩674,960m for the current half and ₩585,010m for the prior half, revenue from major customers accounting for 10% or more of the consolidated entity’s total revenue amounted to ₩499,638m and ₩413,554m respectively.” No names. That is ₩499,638m of ₩674,960m = 74.02% of revenue from customers the note refuses to identify.

But the business-description section, 300 pages away in the same filing, names them with figures: Samsung Electronics ₩419,197m (62.11%), SK hynix ₩80,442m (11.92%), LG Display ₩35,514m (5.26%) — with an explicit footnote that these are consolidated and exclude discontinued operations, so the denominator is the same ₩674,960m.

Add the two customers above 10%: 419,197 + 80,442 = 499,639. The anonymous note says 499,638. A ₩1m rounding difference. The bridge closes, and the anonymous number has names attached to it.

Now the part that a screener will never show you. In H1 2025, BOE Technology Group — China’s largest display panel maker — was 13.15% of Dongjin’s revenue, its second-largest customer. In H1 2026, BOE does not appear on the customer list at all. The company has not lost BOE. It has classified the ten Chinese subsidiaries that served BOE as held for sale, so their revenue moved out of the continuing-operations line that the customer table now measures against.

Check the arithmetic on the 2025 note: Samsung 341,815 + BOE 97,968 + SK hynix 77,877 = 517,660, against the note’s 517,661. The ≥10% club in H1 2025 was three customers, one of them Chinese. In H1 2026 it is two, both Korean.

So the honest way to state the change is not “concentration rose from 69% to 74%.” It is: Dongjin did not become more Samsung-dependent by selling more to Samsung — it became more Samsung-dependent by selling the business that sold to everyone else. And the same ₩419,197m of Samsung revenue is 62.1% of continuing operations but 48.8% of the ₩859,844m Dongjin still consolidates including the discontinued Chinese business (our arithmetic, not the filing’s). Which number you quote depends entirely on which denominator you pick, and both are in the same document.

💵 Where Does Dongjin’s Operating Profit Actually Come From?

All of it comes from one segment: domestic electronic materials earned ₩134,117m of segment operating profit against the group’s total of ₩125,884m, which means every other continuing segment nets out negative.

Three things fall out of that table that the group margin hides.

First, the group is a single-segment business wearing a diversified costume. Domestic electronic materials — effectively the Korean photoresist and wet-chemical operation selling to Samsung, SK hynix and LG Display — produced 106.5% of consolidated operating profit at a 21.7% margin. Everything else, in aggregate, subtracts.

Second, the overseas electronic materials business loses money. ₩93,906m of revenue producing a ₩10,769m operating loss, a −11.5% margin, in the same half that the Korean operation ran at 21.7%. This is the business Dongjin has been funding heavily. The subsidiary schedule states cumulative capital investment of $74.2M in DONGJIN USA and $53.0M in Dongjin Semichem Texas (the latter funded by the former, so the two overlap and should not be added), and itemises $69.6M into Dongjin Sweden AB across 2024–25 — $18.07M in 2024 plus $25.0M and $26.5M in May 2025. Both the US and Swedish entities are classified in the filings’ subsidiary schedule under the same business line, “electronic materials sales”, so both sit inside this segment. What the filings do not do is split the segment result by entity, so we cannot say how the ₩10,769m loss divides between them.

Third, the business being sold is profitable. The Chinese operation earned ₩27,746m of segment operating profit on ₩184,884m of revenue — a 15.0% margin, lower than Korea’s 21.7% but comfortably positive, and it carries ₩340,056m ($253M) of assets held for sale. Selling a 15%-margin business to concentrate on a 21.7%-margin one is a defensible strategy; it is not the same thing as exiting a problem.

📈 Is the Business Getting Better, or Is the Income Statement?

Revenue rose 15.4% and operating profit 39.2% in the first half — a genuinely good result — but net income tripled, and roughly ₩60.6bn of that came from currency movement and marking a share portfolio to market.

Note 26 breaks the ₩61,941m of other non-operating income into its parts, and two lines carry it:

• Foreign exchange, net: +₩32,414m ($24.1M). Realised FX gains ₩13,444m plus translation gains ₩26,512m, less FX losses ₩5,907m and translation losses ₩1,635m. In H1 2025 the same lines netted to −₩25,547m. That single swing is ₩58.0bn.
• Fair-value gains on investment financial assets: +₩28,225m ($21.0M), against ₩792m a year earlier.

Together those two lines are ₩60,639m — 32.7% of pre-tax profit. The second one deserves a sentence of its own. Dongjin’s long-term investment financial assets grew from ₩6,902m at end-2024 to ₩23,366m at end-2025 to ₩55,579m ($41.3M) at 30 June 2026. The filing’s own price-risk note confirms the sensitivity: a 10% move in those holdings changes profit by ₩5,558m — exactly 10% of the balance. A chemicals company’s half-year earnings now move with a share portfolio.

Readers who followed our work on SK hynix, where non-operating income matched operating profit in a single quarter, will recognise the shape. It is not an accounting irregularity — it is IFRS working as designed. It just means the headline net-income growth rate is not a statement about photoresist.

One more trap in the annual numbers. Several data providers show Dongjin’s FY2024 revenue as ₩1,408bn and FY2025 as ₩1,194bn — an apparent 15% decline. It is not a decline. FY2024 as originally filed included the Chinese business; the FY2025 filing restates both years onto a continuing-operations basis, where FY2024 is ₩1,112,813m and FY2025 is ₩1,194,145m, a 7.3% increase. Comparing a restated year against an unrestated one produces a 22-point error in the growth rate.

💰 What Do You Actually Pay for Dongjin Today?

About 21 times last year’s total earnings — but 35 times the earnings that are actually staying with the company, because 38.3% of FY2025 net income came from the Chinese business being sold.

This is the one number where the continuing/discontinued split matters most, so take it slowly. FY2025 consolidated net income was ₩94,092m. Of that, ₩36,060m — 38.3% — was profit from discontinued operations, and only ₩58,032m came from continuing operations. A buyer of the shares today is buying the continuing business; the discontinued profit is being sold along with the subsidiaries that generate it. (Source: DART half-year report to 30 Jun 2026, rcpNo 20260814002890, summary financial information.)

The valuation only means something next to comparable Korean electronic-materials suppliers.

Which column you read reverses the conclusion. On total earnings Dongjin sits mid-pack — cheaper than Solbrain, dearer than Hansol Chemical and ENF. On continuing earnings, the basis a buyer of the shares actually inherits, Dongjin at 35.4× is the most expensive of the four. The three peers have no discontinued operations, so their multiple does not move; Dongjin’s moves by 13.6 turns.

That does not settle anything on its own, and it is not a verdict on the shares. Two things pull the other way: the H1 2026 continuing operating margin of 18.7% is far above the FY2025 14.4% the multiple is computed on, and the continuing figure is depressed by the loss-making overseas segment. But an investor comparing Dongjin to its peers on a screener will see 21.8×, and 21.8× includes earnings the company has agreed to sell.

🏷️ Which “Dongjin” Are You Actually Buying?

Only 005290 on KOSDAQ is listed; every other Dongjin entity in the group — including the holding company that controls it and the foaming-agent business spun out in January 2026 — is unlisted and cannot be bought.

🌏 Is Foreign Money Buying or Selling Dongjin?

It depends entirely on the window: foreigners were net buyers over five sessions and net sellers over twenty, and the two windows point in opposite directions.

Anyone quoting only the five-day column would report that foreign investors are accumulating Dongjin. Over twenty sessions they sold more than three times as many shares as they bought back in the last five, and institutions sold in both windows. The correct description is a partial buy-back of a larger sale, not accumulation. We show two windows on every flow table for exactly this reason — see the Foreign Flow Watch series and our primer on how foreign and institutional flows move Korean stocks.

🏛️ Who Controls Dongjin, and What Do Minority Holders Get?

Chairman Lee Jun-hyuk controls Dongjin Semichem through a two-layer structure: he owns 35.72% of unlisted Dongjin Holdings, which owns 32.78% of the listed company — so roughly 11.7% of the economics carries effective control.

The filings give two different “controlling stake” numbers and both are correct, which is worth understanding rather than picking one:

• 38.53% (19,637,724 shares) — largest shareholder plus all specially related persons, per the half-year report (rcpNo 20260814002890), shareholder section. This includes the Dongjin Scholarship Research Foundation (1,880,000 shares, 3.69%).
• 34.84% (17,757,724 shares) — the group reported in the 5% ownership filing of 8 June 2026, whose named parties are Dongjin Holdings plus Myeongbu Industrial, Ok Yeon-su and Ahn Sang-hee. (Source: DART report of large holdings, rcpNo 20260608000119, filed 8 Jun 2026.)

The two totals differ by 1,880,000 shares — the foundation, which the half-year report counts and the 5% filing does not. 19,637,724 − 17,757,724 = 1,880,000. Neither number is wrong; they answer different regulatory questions, and quoting one as “the controlling stake” without saying which is how a governance figure quietly becomes unreliable.

Dongjin Holdings itself is small: total assets ₩102,702m, equity ₩81,710m ($60.7M) at 31 December 2025. Its 16,706,986 Dongjin Semichem shares are worth roughly $500M at the 4 September 2026 close. It has also pledged shares as collateral — 1,096,651 shares (2.13%) in the prior filing, reduced to 878,159 shares (1.72%) as of 8 June 2026 (same filing, rcpNo 20260608000119). Pledged controlling-shareholder stock is a Korea-specific risk worth tracking: a margin call forces sales the controller did not choose to make.

The buyback that will be cancelled, and the one that will not

On 27 March 2026 Dongjin filed a Value-Up plan under Korea’s corporate-value programme (DART voluntary disclosure, rcpNo 20260327902836). Its commitments:

• FY2025 dividend of ₩650 per share, up ₩450 from the prior year. Total dividends ₩33,400m ($24.8M) versus ₩10,283m ($7.6M) for FY2024 — +224.8%.
• Payout ratio 33.7%, stated explicitly as calculated on FY2025 consolidated net income attributable to owners of the parent. A commitment to hold payout above 25% for three years.
• “We are currently conducting a ₩20bn treasury-share purchase, and when it is complete we will cancel the entire amount. No additional treasury-share acquisition beyond that cancellation has been confirmed, and we plan to focus on cash-dividend-centred shareholder returns going forward.”
• Four investor-relations events a year.

The company did exactly what it promised. The trust ran 8 December 2025 to 13 April 2026, contracted at ₩20,000m and executed at ₩19,999m — 99.99%. On 22 April the board resolved to cancel, and on 29 April 2026 it retired 444,368 shares with a carrying value of ₩21,123m ($15.7M). Shares outstanding fell from 51,414,494 to 50,970,126. The treasury account went to zero.

Then, on 31 July 2026 — four months after saying no further purchase was confirmed — the board approved a second ₩10bn trust with Daishin Securities, to run from 3 August 2026 to 2 February 2027. The reference price stated in that filing is ₩33,150, the previous day’s close, which was also the stock’s lowest close of 2026. (Source: DART material-event report, treasury-share trust signing, rcpNo 20260731000235, filed 31 Jul 2026.)

It did not run to February. The contract was terminated on 31 August 2026, the stated reason being “early termination following completion of the acquisition,” and 234,800 shares were returned from the trust account into the company’s own corporate securities account, where the company now holds them directly. (Source: DART material-event report, treasury-share trust termination, rcpNo 20260831000339, filed 31 Aug 2026.)

One further mechanical point that cuts both ways. The April cancellation raised Dongjin Holdings’ stake from 32.49% to 32.78% without it buying a single share — the half-year report footnotes this explicitly: “the change in ownership ratio is due to the change in shares issued (treasury share cancellation).” Cancellation is genuinely accretive to minorities and simultaneously consolidates family control. In Korea those two facts travel together, which is part of why the Korea Discount is a governance question before it is a valuation one.

Finally, related-party purchasing. The filing states plainly that Dongjin Advanced Materials — one of the domestic suppliers of the solvents, PAC and resin that go into photoresist — “is a specially related person of the largest shareholder.” Total purchases from all related parties in H1 2026 were ₩41,684m ($31.0M), against ₩387,678m of electronic-materials raw materials bought across the domestic electronic-materials, domestic refined-oil and overseas electronic-materials segments — roughly 10.8%. The filing’s counterparty breakdown is a nested table we could not attribute line-by-line with confidence, so we quote only the total. Disclosed, audited, and not large enough on its own to be alarming — but it is money flowing to entities the controlling family owns more of than it owns of the listed company.

📚 Lingo Check

🎯 Why It Matters for K-Export Stars

Every finding in this article that changes the conclusion exists only in Korean-language regulatory filings — none of these documents has an English version, and Dongjin does not have an English Wikipedia page.

Every fact in this article that changes the conclusion came from a Korean-language regulatory filing with no English equivalent — the named customer table, the segment operating-profit split, the note reconciling ₩499,638m to two companies, the restatement that turns a 15% revenue decline into a 7.3% increase, and the one line stating the second buyback goes to employees rather than to cancellation. None of it is hidden. It is simply not translated.

Dongjin is the clearest case yet for why this matters. It is a supplier, not a headline name: it will never appear in a Yonhap wire story the way SK hynix or Samsung Electronics do. But it sells into both of them, which makes it a leveraged read on Korea’s chip cycle — with all the concentration risk that implies, and a governance structure you can only assess by reading the filings. The same pattern we mapped at Samsung Electro-Mechanics: the component makers carry the cycle’s beta without its brand.

This company is one of ten in our K-Semiconductor Filing Map — a supply-chain map where every cell is marked by what the filing itself supports, and each one links to the receipt number it came from.

Disclaimer: This article is for informational and educational purposes only and does not constitute investment advice, a recommendation, or an offer to buy or sell any security. All figures are sourced from DART regulatory filings and Korea Exchange settled market data as dated inline, and may change. Prices and valuation ratios are point-in-time as of 4 September 2026. Do your own research and consult a licensed financial adviser before making any investment decision.